Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Validation setup #13

Merged
merged 13 commits into from
Jan 4, 2019
10 changes: 10 additions & 0 deletions research/README.md
Original file line number Diff line number Diff line change
@@ -1,3 +1,13 @@
# Research

**Owner**: [Purshottam Shivraj](https://github.com/pshivraj)

Current Research work has been around exploring and setting up validation setup for clomask using pre-defined validation images using [matterport](https://github.com/matterport/Mask_RCNN) implimentation.

Current validation setup is measured against mAP which is defined as average of the maximum precisions at different recall values and is benchmarked at **0.5676** based on the validation process explored in the [clomask_notebook](https://github.com/havanagrawal/clomask/blob/validation_setup/research/clomask_notebook.ipynb).

## Current limitation of model

### Matterport(default config)
* Slant bottles are not detected.
* Some bottle images having connected lables between them are merged as one causing single mask for two different bottles, throwing off mAP.
535 changes: 535 additions & 0 deletions research/clomask_notebook.ipynb

Large diffs are not rendered by default.